Aims: To evaluate the occurrence and abundance of phages that carry the stx1 and stx2 gene in water samples of different quality.
Methods and Results: Phages growing on the Shiga toxin-negative Escherichia coli O157:H7 (ATCC 43888) strain were enumerated by a plaque assay in concentrated raw and treated waste water samples and river water samples. Plaques were investigated for the presence of stx1 and stx2 genes by a multiplex/nested PCR procedure. An overall number of 805 plaques were tested for the presence of stx-carrying phages. Stx genes could be demonstrated in 2% (stx1) and 16% (stx2) of the plaques. Stx-phages were eliminated with approximately the same efficiency in comparison with somatic coliphages during the waste water treatment process.
Conclusions: Due to the low numbers of phages carrying the stx genes 1 and 2 in treated waste water and river water, the dilution and inactivation of host bacteria and the unsuitable conditions for the transduction of host organisms in aquatic environments, it is difficult to derive from the data the direct evidence for a public health problem.
Significance and Impact of the Study: The results show the quantitative occurrence of stx-carrying phages in waste and river water and confirm the frequent circulation of these viruses in the aquatic environment.
Shiga toxin-producing Escherichia coli (STEC) are associated with the occurrence of diarrhoea, haemorrhagic enterocolitis and the haemolytic-uremic syndrome in the human population (Tarr 1995). Escherichia coli serotype O157:H7 is the main causative agent of these diseases although other serotypes of E. coli and other enterobacteria species are also involved in the pathogenesis (Paton and Paton 1998). The enterohaemorrhagic E. coli serotypes produce at least two different Shiga toxins (Stx1 and Stx2) which are encoded by phages (O'Brien et al. 1984). Stx-encoding bacteriophages investigated so far consist of double-stranded DNA and showed lambdoid morphology (Herold et al. 2004). The genetic information of the Shiga toxins can be transferred from phages to the DNA of non-Shiga toxin hosts belonging to different species of Enterobacteriaceae by transduction (Beutin et al. 1999; Schmidt et al. 1999). Most of the manifestations of the STEC diseases are thought to be a result of the effect of Stx although the correlation of Stx2 with severe disease in humans is stronger in comparison with other Shiga toxins (Boerlin et al. 1999). Besides the two principal members of the Stx family, Stx1 and Stx2, several variations of these toxins have been reported in the last years showing differences in virulence and distribution in the investigated host populations (Thorpe et al. 2002). Most of the reported outbreaks of STEC infections are associated with the consumption of contaminated foods (Nataro and Kaper 1998) but waterborne infections are described (Friedman et al. 1999; Chalmers et al. 2000; Bopp et al. 2003). The role of phages in the transfer of stx genes within the aquatic environment is not clear to date. The first study to detect phages encoding the stx2 gene has demonstrated high levels of such phages in raw waste water samples (Muniesa and Jofre 1998) and was confirmed by the investigation of sewage samples collected in other countries (Muniesa and Jofre 2000). Muniesa et al. (1999) demonstrated in in situ inactivation experiments a remarkable persistence of phages carrying the stx2 gene. Nevertheless, no phages infected to two E. coli O157:H7 strains could be detected in supernatants from the secondary settling tank and the chlorinated effluent of a waste water treatment plant (Tanji et al. 2003). Little is known about the prevalence of phages carrying stx genes other than stx2 in water samples or the occurrence of stx-phages in surface waters receiving domestic waste water. Here, we report data on the abundance of phages carrying stx1 and stx2 genes in different water systems by a combination of cultural and molecular methods to evaluate the potential of a possible stx gene transfer from stx phages to non-stx bacteria under natural conditions.
Materials and methods
Water samples and sample preparation
Raw and treated waste water samples were obtained from two distinct plants with both activated sludge treatment and chemical phosphate removal in Saxony, Germany. The plants treat urban sewage with a capacity of 5000 m3 day−1 (plant 1) and of 120 000 m3 day−1 (plant 2) respectively. The list of registered waste water dischargers let assume that there is no strong influence of faeces of animal origin on both waste water systems. River Elbe was sampled in the area of the city of Dresden. The samples were transported in coolers (4°C) and processed within 12 h after sampling. Either 24 ml (raw waste water) or 325 ml (treated waste water and river water) of the water samples were concentrated by ultracentrifugation (120 000 g, 3 h, 4°C), the pellets were resuspended in 5–10 ml beef extract (10%, pH 7·5; Becton Dickinson, Sparks, MD, USA) and stirred (1 h) at room temperature. Particulate matter and bacteria were removed by low-speed centrifugation (3000 g, 15 min, 4°C) and filtration (pore size: 0·22 μm, pretreated with 3% beef extract to reduce phage adsorption to the filter). The filtrates were subjected to the enumeration of coliphages.
Bacterial strains and detection of somatic coliphages and coliphages growing on E. coli ATCC 43888
Escherichia coli C (ATCC 13706) was used for the enumeration of somatic coliphages. Escherichia coli O157:H7 EDL933 (ATCC 43895) which produces Stx1 and Stx2 was used as positive control for the amplification of stx genes. Escherichia coli O157:H7 ATCC 43888 (provided by M. Muniesa, University of Barcelona, Spain) which does not produce either Stx1 or Stx2 and does not possess these genes was the host strain for the determination of phages carrying the different stx genes in water samples. All strains were grown in modified Scholten's broth or on modified Scholten's agar (Anonymous 2000). The E. coli strains O157:H7 EDL 933 and E. coli O157:H7 ATCC 43888 were incubated for 5 h at 37°C, centrifuged and the sediment was resuspended in 2 ml of phosphate-buffered saline. A volume of 200 μl of the suspension was used for the extraction of bacterial DNA. DNA extraction of E. coli ATCC 43888 was repeated at the end of the study to exclude the acquisition of stx genes in the host strain via a laboratory contamination. Somatic coliphages growing on E. coli ATCC 13706 and phages infecting E. coli ATCC 43888 were enumerated by the double layer agar technique according to the instructions of ISO 10705-2 (Anonymous 2000) and the phage concentrations were expressed as plaque forming units (PFU).
Polymerase chain reaction for the detection of stx genes
Plaques in E. coli ATCC 43888 were picked with sterile toothpicker and directly amplified in a combination of a multiplex/nested polymerase chain reaction (PCR) approach. Primers (ThermoHybaid, Ulm, Germany) for the first amplification and the nested PCR are summarized in Table 1. The primer pairs for the detection of the stx2 gene were as described by Muniesa and Jofre (1998). Primers for the amplification of the stx1 gene were derived from the published sequence of the subunit A of the stx1 gene (EMBL accession number M19473). First amplifications of both stx1 and stx2 gene in one reaction were carried out in a final volume of 25 μl containing 2·5 μl 10x PCR buffer (Applied Biosystems, Foster City, CA, USA), 2·5 μl dNTP (200 mmol l−1 each; Hybaid-AGS, Heidelberg, Germany), 0·125 μl AmpliTaq-Gold (Applied Biosystems) and 1 μl of each primer (10 pmol). The reaction mix was heated for 10 min at 94°C and the DNA was amplified by 35 cycles using the following settings: 60 s at 94°C, 60 s at 56°C, 60 s at 72°C and final 10-min extension at 72°C. A volume of 1 μl portion of the first-round PCR was used for the nested PCR, which was carried out separately for the two genes. The amplification programme was as described but with an annealing temperature of 50°C. Genomic DNA from the E. coli O157:H7 strains EDL 933 and ATCC 43888 was extracted with the QIAamp tissue mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions and 1 μl DNA of E. coli O157:H7 EDL933 and 1 μl water was used as positive and negative control in both reactions. PCR products were stained in ethidium bromide and analysed on a 1·5% agarose gel. Randomly selected products of the nested PCR were treated with the QIAquick PCR purification kit (Qiagen) and sequencing was performed using the BigDye terminator cycle sequencing kit and the DNA sequencer ABI Prism 377 (Perkin-Elmer Applied Biosystems) according to the manufacturer's instruction.
Table 1. PCR primer sets used for amplification of stx genes
The concentrations of somatic coliphages growing on E. coli C, the concentrations of phages infecting E. coli ATCC 43888 and the results of the detection of stx genes in the plaques obtained from water samples of the different sites are summarized in Table 2. Mean influent concentrations of somatic coliphages of 1436 PFU ml−1 (waste water treatment plant 1) and 1877 PFU ml−1 (plant 2) were reduced during the treatment process to 2·5 and 43 PFU ml−1 respectively. The mean concentrations of phages infecting E. coli ATCC 43888 in raw waste water ranged between 16·9 PFU ml−1 (plant 1) and 37·2 PFU ml−1 (plant 2) whereas in treated waste waters and in the river Elbe mean concentrations of 0·03 PFU ml−1 (plant 1), 1·8 PFU ml−1 (plant 2), and 0·2 PFU ml−1 (river water) were determined. The mean percentage of phages infecting E. coli ATCC 43888 varied between 1% (raw waste water of treatment 1), 2% (raw waste water of treatment 2), 1% (effluent of treatment 1), 4% (effluent of treatment 2) and 11% (river water) of the concentration of somatic coliphages. The plaque morphology of phages infecting E. coli 43888 was in all cases small, round and transparent. A total of 805 plaques developed in E. coli ATCC 43888 from raw and treated waste water samples of two different plants and samples from the river Elbe were analysed for the presence of stx1 and stx2 genes. Figure 1 shows an example for the typical amplification products of the nested PCR of stx1 (329 bp) and stx2 (169 bp) gene obtained from plaque material of a waste water sample. In raw waste water samples the stx1 gene alone was present in 3 (plant 1) and 5 plaques (plant 2) corresponding to 1% of all investigated plaques. No stx1 genes could be detected in plaques developed from treated waste water samples of both plants. Six of 142 plaques (4%) developed from river water samples showed a stx1-positive PCR. Stx1 genes could be found in three of the seven investigated surface water samples only with rates ranging between 2% and 17%. In raw water samples 52 plaques (19% of all investigated plaques) of plant 1 and 55 plaques (16%) of plant 2 were positive only for stx2. In one plaque (11%) of treated waste water from plant 1 the stx2 gene was present. In none of the investigated 30 plaques from treated waste water samples of plant 2 the stx2 gene could be amplified. In samples of the river Elbe the investigation of 28 plaques (20% of all investigated plaques) resulted in amplification products specific for the stx2 gene. The amplification of stx2 genes was successful in three surface water samples demonstrating rates between 5% and 40% of stx2-positive plaques. Stx1 and stx2 genes in combination were present in 1% of plaques of raw waste water samples of both treatment plants and the river water samples.
Table 2. Somatic coliphages infecting Escherichia coli C, phages growing in E. coli ATCC 43888 and phages carrying the stx1 and stx2 genes in water samples
Number of samples
Mean concentration of somatic coliphages in PFU ml−1 (min–max)
Mean concentration of phages infecting E. coli ATCC 43888 in PFU ml−1 (min–max)
Overall number of plaques picked for detection of stx genes
Number of stx1-positive plaques (% of picked plaques)
Number of stx2-positive plaques (% of picked plaques)
Number of stx1- and stx2-positive plaques (% of picked plaques)
Waste water treatment plant 1
Waste water treatment plant 2
Six of 14 PCR products of stx1 gene (43%) and 35 of 133 PCR products of stx2 gene (26%) were sequenced. They showed the identical sequence as the stx1 and stx2 genes deposited in the EMBL GenBank (accession numbers: M19473 and X07865).
Although no information was available concerning the possible part of the population of the stx-phages showing replication without plaque formation, the direct characterization of phages carrying the stx1 and stx2 genes from plaques grown on E. coli O157:H7 ATCC 43888 allows a precise quantification of the occurrence of these phages in a concentrated water sample. Furthermore, this study approach has the advantage to detect infectious phages in a simple test system resulting in PCR products, which can be used for sequence analysis. The sequencing of the selected PCR products obtained from different water samples confirmed that the investigated phages were stx-phages and resulted in uniform sequences for the amplified parts of both of the stx genes. This result is an indication for the use of faster hybridization techniques to characterize the plaques.
The evaluation of the concentration of phages infected E. coli O157:H7 in water samples showed to be difficult. In raw waste water samples Tanji et al. (2003) found remarkable seasonal differences of the phage concentrations exceeding more than two log units and a concentration of phages infected E. coli O157:H7 EDL933, which was almost one-tenth of that of E. coli O157:H7 ATCC 43888 phages. Muniesa and Jofre (2000) detected stx-phages in 15 of 33 raw waste water samples which were collected in 10 different countries. These samples were characterized by varying sources of contamination and a mean ratio between somatic coliphages and those infecting E. coli O157:H7 ATCC 43888 of 20 : 1 could be calculated. The average concentration of phages infecting E. coli ATCC 43888 was 6·1 × 102 ml−1 with a range of <10 and 9·7 × 103 PFU ml−1 indicating different sources of waste water contamination and probably different epidemiological situations in the investigated areas. In our study, the mean concentrations of phages infecting E. coli O157:H7 ATCC 43888 ranged between around 1% (waste water of treatment plant 1) and 10% (river Elbe) of values calculated for the somatic coliphages suggesting differences concerning the input and/or the survival of both phage groups in the investigated water resources. Muniesa and Jofre (1998) estimated the numbers of phages carrying the stx2 gene by using the most probable number technique between 100 and 1000 per 100 ml raw waste water representing around 1% of the phages infecting E. coli O157:H7 ATCC 43888. With the data of our work mean concentrations of around 34 (stx1) and 340 PFU (stx2) per 100 ml raw waste water were calculated showing a similar range of contamination with stx2-phages in both studies and the 10 times higher abundance of stx2-phages in comparison with phages containing the stx1 gene in domestic waste water. The results of Tanji et al. (2003) indicated a reduction of phages infecting E. coli ATCC 43888 in a waste water treatment plant with nutrient removal of more than three logs. We estimated in plant 1 a mean reduction of both phage populations of 99·8%, whereas 97·7% (somatic coliphages) and 95·2% (phages infecting E. coli O157:H7) were removed in plant 2 suggesting a comparable efficiency of the water treatment on both phage populations. The faeces of humans can be assumed as source of phages carrying the stx genes in the investigated waste water samples. Besides the effluents of sewage treatment plants diffuse contaminations of agricultural origin in the intensively used catchment area of the river Elbe might be an influencing factor on the composition of phages carrying the stx1 gene in the river samples in comparison with the inflows of both treatment plants. STEC were detected in the faeces of numerous animal species (Duffy 2003; Caprioli et al. 2005) including some non-O157:H7 serotypes which produce predominately Stx1 (Johnson et al. 1996). The concentrations of somatic coliphages in samples of the river Elbe were typical for surface waters with ‘intermediate’ pollution with human and non-human faeces (Araujo et al. 1997). The mean concentrations of phages carrying the stx genes in the positive river water samples can be estimated around one (stx1) and seven (stx2) per 100 ml water. The detection of phages carrying the stx1 and stx2 genes in surface water proved that the occurrence of these phages are not restricted to raw waste water (Muniesa et al. 1999). Inactivation experiments confirmed that the phages containing the stx2 gene persist in water longer than their host bacteria (Muniesa et al. 1999) suggesting a reservoir of these genes in the phage populations of waste and surface waters with the potential for the horizontal transfer of stx genes. Unfortunately, no data are available concerning differences between the stability of stx1- and stx2-phages which could be a further reason for the higher abundance of phages carrying the stx1 gene in river water samples.
Because of the low numbers of phages carrying the stx genes 1 and 2 in treated waste water and river water, the dilution and inactivation of host bacteria and the unsuitable conditions for the transduction of host organisms in aquatic environments it is difficult to derive from the data the direct evidence for a public health problem. Further experiments are necessary to define the conditions for the successful transmission of these genes in nature and to estimate the role of stx-phages in the epidemiology of STEC infections via the water route.
This study was supported by the Bundesministerium für Bildung und Forschung (02WA0176).